Gas mixing and distribution

We have investigated gas mixtures in search of an environmentally friendly and non-combustible mixture that provides high detection efficiency and stable RPC operation [76]. We compared 16 different mixtures with butane concentrations of 4, 8, 12, and 25 % and argon concentrations of 20, 25, 30, and 35 % with the balance of the gas being HFC-134a. The RPC performance in terms of efficiency, dark current, singles rate, and timing resolution was compared at an operating point 200 V/mm above the knee of the efficiency plateau curve. The butane reduces the presence of after-pulses by absorbing photons from the initial discharge. The non-flammable limit for the butane is about 12 % at the mixing ratio of 1:1 for the argon and HFC-134a. We found very little difference between flammable and non-flammable mixtures and have chosen a non-combustible mixture of 62 % HFC-134a, 30 % argon, and 8 % butane-silver. Table lists some basic physical parameters of these gases. Butane-silver is a mixture of approximately 70 % n-butane and 30 % iso-butane. The cost of butane-silver is one tenth of the cost of 99.5 % pure iso-butane.

Table: Physical parameters of the gases used in KLM.

Gas	Symbol	Mol. weight	Density (g/$l$)
Argon	Ar	39.95	1.784 (0 $^o$C, 1atm)
Butane-silver	C$_4$H	58.12	2.6 (0 $^o$C, 1atm)
HFC-134a	CH$_2$FCF$_3$	102.0	4.5

Two separate banks of bottles are arranged for each type of gas. When one side becomes empty, the supply line automatically switches to the other. Tank quantities are measured by weight for butane and HFC-134a and by pressure for argon. A diagram of the mixing system is shown in Fig. . The three gases are sent to MKS model 1179A mass flow controllers for mixing in the appropriate ratios. Four gas mixing systems are used separately for the inner RPCs in the barrel super-layers, the barrel outer RPCs, the end-cap inner RPCs, and the end-cap outer RPCs. The flow rates from the mass flow controllers are monitored via a network connection and the high voltage is automatically lowered if a deviation from the desired flow rate is detected. During normal operation, we flow a total of 4.5 $l$/min, which corresponds to approximately one volume change per day.

Figure: KLM gas mixing system.

The gas distribution system is designed to provide an independent gas supply to each RPC in a super-layer. Therefore, if one supply line fails for any reason, the other RPC in the same super-layer will still be operational. To ensure uniform distribution of the flow without the need for tedious adjustments, a "flow resistor" was inserted in series upstream of each RPC. These devices are 10 cm-long stainless-steel tubes with an inner diameter of 254 $\mu$m. The flow impedance of the tubes is about ten times larger than that of an RPC layer. Thus the flow rate is determined by the flow resistor (uniform to about 15 %) and almost independent of variations in the flow resistance of individual RPCs.
The exhaust system has an active control of the exhaust pressure and relief bubblers at various points in the system to prevent any overpressure situations. Tests indicated that epoxy joints between the glass plates and the internal spacers begin to detach when a barrel RPC is pressurized above 50 mmAq. For safety reasons, the exhaust gas is dumped outside the experimental hall through a 20 m vertical exhaust line. Due to the large density of the mixed gas a suction pump system is used.